14 October 2019
Laundry Robotics has been using Fizyr’s vision software in their robots, since 2017. Robots can work 24/7 and are able to fold up to 720 towel per hour.
Project number: E!113339
Our software enables robotic systems to pick a large variety of items and detect outliers, whether the robot has seen them before or not. Thanks to its deep-learning algorithms, it proposes over 100 grasp poses each second on unknown objects.
This way, automated picking solutions gain high level of autonomy to deal with unstructured environments and handle complex conditions such as deformation, damage, wrinkles, texts or any other varying pattern on the items and packages.
In the summer of 2017, Fizyr solved a computer vision problem that scientists have been trying to crack for many years. We have trained a robust neural network to detect towel corners from bulk. A combination of deep learning and unique heuristics algorithms allow us to fully automate the towel folding process.
Since then, Laundry Robotics has applied our product to automate the folding process at laundry services worldwide. Based on this success, we continued applying our software product to enable the automated picking of apparel, bags, luggage and other deformable items. You can explore more on Fizyr’s software applications here.